University of South Carolina University of South Carolina

Scholar Commons Scholar Commons

Theses and Dissertations

Fall 2019

Predicting Change in Autism Symptomatology in Young Children Predicting Change in Autism Symptomatology in Young Children

at Risk for Autism Spectrum Disorder: Fragile X Syndrome, Down at Risk for Autism Spectrum Disorder: Fragile X Syndrome, Down

Syndrome and Non-Syndromic ASD Syndrome and Non-Syndromic ASD

Kelly Elizabeth Caravella

Follow this and additional works at: https://scholarcommons.sc.edu/etd

Part of the Community Psychology Commons

Recommended Citation Recommended Citation Caravella, K. E.(2019). Predicting Change in Autism Symptomatology in Young Children at Risk for Autism Spectrum Disorder: Fragile X Syndrome, Down Syndrome and Non-Syndromic ASD. (Doctoral dissertation). Retrieved from https://scholarcommons.sc.edu/etd/5606

This Open Access Dissertation is brought to you by Scholar Commons. It has been accepted for inclusion in Theses and Dissertations by an authorized administrator of Scholar Commons. For more information, please contact digres@mailbox.sc.edu.

PREDICTING CHANGE IN AUTISM SYMPTOMATOLOGY IN YOUNG CHILDREN AT RISK FOR AUTISM SPECTRUM DISORDER: FRAGILE X SYNDROME, DOWN

SYNDROME AND NON-SYNDROMIC ASD

by

Kelly Elizabeth Caravella

Bachelor of Arts Tulane University, 2010

Master of Arts

University of South Carolina, 2016

Submitted in Partial Fulfillment of the Requirements

For the Degree of Doctor of Philosophy in

Clinical-Community Psychology

College of Arts and Sciences

University of South Carolina

2019

Accepted by:

Jane Roberts, Major Professor

Mark Weist, Committee Member

Kimberly Hills, Committee Member

Jessica Klusek, Committee Member

Cheryl L. Addy, Vice Provost and Dean of the Graduate School

ii

© Copyright by Kelly Elizabeth Caravella, 2019 All Rights Reserved.

iii

ABSTRACT

This dissertation is comprised of two manuscripts which examine the longitudinal

development of autism symptomatology in young children at risk for developing autism

spectrum disorder (ASD); individuals with Fragile x syndrome (FXS) and Down

syndrome (DS). The first study is a within group analysis of the longitudinal development

of ASD symptomatology in young children with FXS, and how diagnostic stability,

language and non-verbal cognitive functioning may predict these trajectories. This paper

provides insight into ASD diagnostic stability patterns within FXS, and how symptoms

change over time across these groups. The second paper will extend this work by

presenting a pilot cross-syndrome comparison, which includes young children with FXS,

DS and non-syndromic ASD to examine hypothesis of unique ASD symptom trajectories

between the disorders. This pilot examination will set the stage for future larger scale

cross syndrome comparisons to target these important questions. Together, the findings

presented here inform the theoretical understanding of ASD within FXS, as well as

support clinical recommendations for screening, monitoring and diagnosing ASD within

FXS.

iv

TABLE OF CONTENTS

Abstract .............................................................................................................................. iii

List of Tables ...................................................................................................................... v

List of Figures .................................................................................................................... vi

List of Abbreviations ........................................................................................................ vii

Chapter 1: Predicting Change in Autism Symptomatology in Young Children with Fragile X Syndrome ............................................................................................................ 1

Chapter 2: Cross-Syndrome Contrasts of Autism Symptoms in Young Children: Fragile X, Down Syndrome and Non-Syndromic Autism ............................................................ 32

References ......................................................................................................................... 54

v

LIST OF TABLES

Table 1.1 Participant Demographics for FXS Males ........................................................ 29

Table 1.2 Results of a taxonomy of multilevel models for change in BOSCC scores over time ................................................................................................................................... 30 Table 2.1 Mean BOSCC scores by group across time ..................................................... 52

vi

LIST OF FIGURES

Figure 1.1. Uncontrolled effects of diagnostic stability group on BOSCC scores over time ................................................................................................. 31 Figure 2.1. Individual spaghetti plots of BOSCC total score trajectories between groups ................................................................................................................ 53

vii

LIST OF ABBREVIATIONS

ADI ......................................................................................... Autism Diagnostic Interview

ADOS .................................................................. Autism Diagnostic Observation Schedule

ASD ............................................................................................ Autism Spectrum Disorder

ASIB ................................................................................................... Autism Infant Sibling

BOSCC ............................................. Brief Observation of Social Communication Change

CBE .................................................................................................... Clinical Best Estimate

DS .............................................................................................................. Down Syndrome

ELC ................................................................................ Mullen Early Learning Composite

FXS ....................................................................................................... Fragile X Syndrome

ID ........................................................................................................ Intellectual Disability

NSASD ............................................................ Non-syndromic Autism Spectrum Disorder

SCQ ........................................................................... Social Communication Questionnaire

1

CHAPTER 1

PREDICTING CHANGE IN AUTISM SYMPTOMATOLOGY IN YOUNG CHILDREN WITH FRAGILE X SYNDROME

Autism Spectrum Disorder (ASD) is a behaviorally defined disorder characterized

by core impairments in social communication and the presence of atypical restricted

interests and repetitive behaviors (American Psychiatric Asssociation, 2013). ASD is a

heterogeneous disorder, with cognitive, language and adaptive impairments ranging

significantly between individuals. This significant heterogeneity has led the field to shift

focus to the identification of possible distinct phenotypes within the autism spectrum,

labeled the “autisms” by some (Geschwind & Levitt, 2007). Early work aimed at

identifying mechanisms targeted genetic factors given the established heritability present

in twin studies (Geschwind, 2011). However, genetic biomarkers underlying ASD are

complicated with the identification of over 1000 potential risk genes (Vorstman et al.,

2013), and at least 4 single gene disorders with significant relationships to ASD

(Geschwind, 2011).

Fragile X syndrome (FXS) has the highest penetrance of any single gene disorder

implicated in ASD, with between 60-75% of individuals meeting criteria for ASD (Harris

et al., 2008; Klusek, Martin, & Losh, 2014) and upwards of 90% exhibiting at least one

symptom (Bailey, Hatton, Mesibov, Ament, & Skinner, 2000). Males with FXS and ASD

also exhibit a more homogenous behavioral phenotype (e.g. intellectual disability,

language delay) than individuals with non-syndromic ASD (nsASD), making it an ideal

2

model for the study of the development of ASD. By examining a more homogenous

group of individuals with ASD, targeted treatments, similar to the precision medicine

approach in the physical health field, may be achieved.

FXS is caused by an expansion on the FMR1 gene of the CGG repeat sequence on

the X chromosome. In the typical population, average CCG repeat length is less than 55,

with expansions of 55-200 repeats considered the fragile X premutation. When the

expansion exceeds 200 repeats, a reduction in Fragile X Mental Retardation Protein

(FMRP) ensues. FMRP is essential for cognitive development (Garber, Visootsak, &

Warren, 2008). Males and females with FXS are differentially impacted by the disorder

given the protective factor afforded to females by the presence of a second, typically

unaffected, X chromosome. Therefore, females with FXS are often less severely impaired

across a range of domains than males (Hagerman & Hagerman, 2002).

Autism Features in FXS

Core features of ASD including social impairments and restricted and repetitive

behaviors are generally accepted as common behavioral features in FXS, however

questions remain as to the whether or not these features may occur on a spectrum within

individuals with FXS or if they represent a true comorbidity of ASD. These questions

have been part of an ongoing discussion in the literature for the last 30 years in an

attempt to explain the elevated proportion of males with FXS and ASD. Initial theories to

explain the elevated features of ASD in FXS (Cohen et al., 1991) proposed possible

pathways driven by biological markers of FXS (e.g. FMRP) or secondary consequences

of almost universal intellectual disability. Recent work has sought to provide clarity on

this issue by directly testing these relationships. Hall and colleagues examined the

3

relationships between parent reported autism symptoms and biological markers (i.e.

FMRP) and cognitive functioning in a sample of 120 individuals with FXS 5-25 years of

age. Cognitive functioning, but not FMRP, was related to autism symptomatology with

lower cognitive functioning associated with greater parent reported autism

symptomatology (Hall, Lightbody, Hirt, Rezvani, & Reiss, 2010). The null relationship

between FMRP and autism symptomatology is consistent with previous work reporting

similar findings (Donald B Bailey Jr, Hatton, Skinner, & Mesibov, 2001; Rogers,

Wehner, & Hagerman, 2001).

Others have approached the question of unique and shared ASD features in FXS

and non-syndromic ASD through a comparative behavioral lens. This approach explores

whether behavioral features across individuals with nsASD and FXS with comorbid

autism (FXS+ASD) are parallel, which would provide evidence that ASD in both groups

could arise from a similar etiology. Findings from this body of work suggest a

developmental effect in which autism symptomatology is remarkably similar between 21-

48 month olds with FXS+ASD and nsASD (Rogers et al., 2001), with a unique

behavioral phenotype of strengths in the quality and frequency of social overtures in

those with FXS+ASD, that is not seen in nsASD, becoming evident over time (McDuffie,

Thurman, Hagerman, & Abbeduto, 2014; Wolff et al., 2012). These findings have led

some researchers to hypothesize that the features of ASD seen in FXS do not arise from

the same mechanistic underpinnings (Abbeduto, McDuffie, & Thurman, 2014; Hall et al.,

2010). This is supported by evidence of structural brain abnormalities in young children

with nsASD that differ from those with FXS+ASD despite similar behavioral features of

ASD across the two groups (Hazlett et al., 2009).

4

Despite possible mechanistic differences giving rise to shared features of ASD

across groups, consistency exists in findings that individuals with FXS+ASD are more

impaired across a range of domains that those with FXS without ASD (FXS-noASD)

including cognition (Bailey, Hatton, Tassone, Skinner, & Taylor, 2001; Hogan et al.,

2017; Kaufmann et al., 2004), adaptive behavior (Caravella & Roberts, 2017; Hahn,

Brady, Warren, & Fleming, 2015), language (Klusek et al., 2014), motor skills (Bailey et

al., 2000) and social approach behaviors (Roberts, Weisenfeld, Hatton, Heath, &

Kaufmann, 2007). Additionally, children with FXS+ASD receive more hours of speech-

language and behavioral therapy than those without an ASD diagnosis (Martin et al.,

2013). Therefore, identifying ASD features in FXS is essential for providing access to the

appropriate dose and type of intervention required to reduce functional impairment.

Identifying predictors of ASD in young children with FXS is essential towards efforts to

improve treatment with several features emerging as salient. Sex is one of the most well-

established predictors of ASD in FXS, with males being significantly more likely to show

impairing symptoms of ASD as well as meet diagnostic criteria (Hall et al., 2010; Klusek

et al., 2014). Cognitive level and language ability have also been identified to predict

autism symptomatology in FXS beginning in infancy (Hogan et al., 2017) and

maintaining through childhood and early adolescence (Klusek et al., 2014). Greater

pragmatic language impairments have also been found as predictors of increased autism

symptom severity (Lee, Martin, Berry-Kravis, & Losh, 2016).

Stability of Autism Symptomatology in nsASD and FXS

Stability of a positive ASD diagnosis in non-syndromic populations has been

reported to be between 53-100% (Woolfenden, Sarkozy, Ridley, & Williams, 2012). In

5

community samples, autism is generally stable (~90%) by 2 years of age when using a

clinical best estimate (CBE) procedure informed by gold standard measures of autism

symptomatology (e.g. Autism Diagnostic Observation Schedule), developmental

measures and clinical judgment (Chawarska, Klin, Paul, & Volkmar, 2007). Instability in

ASD diagnoses presents in individuals who display milder features or inconclusive

presentations at 2 years old (Guthrie, Swineford, Nottke, & Wetherby, 2013). In a sample

of 418 high risk infant siblings of children with ASD assessed at 18, 24 and 36 months,

diagnostic stability at 24 months was 82% (Ozonoff et al., 2015). However, when the

sample was followed to age 3, 41% of the sample who met diagnostic criteria for ASD at

3 years old were determined to have been false negatives at the 24 month visit (Ozonoff

et al., 2015). Early false negatives were found to be behaviorally atypical at 24 months

(e.g., developmental delays, presence of autism symptomatology), however, their

presentation was not yet clear enough to meet diagnostic criteria for ASD. Taken

together, this research suggests that diagnostic stability in toddlers with nsASD is

dependent on clarity of the clinical presentation, which may not occur until 3 years of age

or later, highlighting the need for the monitoring of autism symptomatology into the third

year of life and beyond. Also, evidence suggests when instability occurs, it more often

represents false negatives at a young age than false positives.

Within the FXS literature, less is known about diagnostic stability across early

childhood with no published studies using a CBE process to inform diagnosis of ASD.

Research into the stability of ASD diagnoses and symptomatology in FXS to date has

largely relied on diagnostic categorizations determined by measure cutoffs, instead of

clinician judgment. Hernandes and colleagues (2009) used the ADI-R, a parent interview

6

about lifetime autism symptomatology, to inform diagnoses of ASD in a sample of 56

males with FXS, who had a mean age of 56 months at baseline. They reported general

diagnostic stability (68%) when participants were tested over the course of 3 years, with

far less stability in individuals with milder symptomatology (21%). At the symptom

level, the two diagnostic groups became less differentiated over the course of the study;

participants in the FXS+ASD group evidenced a reduction in autism symptomatology

while those with FXS-noASD, showed an increase in autism symptoms.

In a slightly older sample of 65 children with FXS (31 male, 34 female), ages 8-

11 years, Lee and colleagues (Lee et al., 2016) used the ADOS to report diagnostic

stability and symptom change over time across 2 visits that were on average 2.5 years

apart. Across the 2.5 years, the percentage of male participants meeting criteria for ASD

increased from 54% to 80%, with most of the increasing impairment found in the social

communication domain. Rates of ASD were remarkably stable in the female participants,

which were reported to be 41.5% at time 1, and 41.2% at time 2. The study did not report

if any participants meeting criteria for ASD at time 1 failed to meet criteria at time 2, so

exact stability estimates are unknown.

In the largest longitudinal study to date (n=116), Hatton and colleagues (2006)

examined Childhood Autism Rating Scales (CARS-2, Schopler, Reichler, & Renner,

2002) total scores in a sample of 1.5 to 15 year olds. The CARS-2 is a clinician rated

measure that is completed after an observation and interactions with a child to rate

symptoms of ASD. It provides clinical categorizations into three groups based on raw

score totals; non-autistic, mild-to-moderate autistic behaviors and severe autistic

behaviors. The CARS-2 is considered to be a screener for ASD symptoms and is not a

7

diagnostic tool unless used in conjunction with other measures. When examining raw

scores over time, the authors reported an average increase of 2 points over 10 years.

Using the clinical categorizations provided by the CARS-2, the authors reported that

about 54% of the participants maintained their diagnostic category throughout the course

of the study, with 13% evidencing decreasing symptomatology and 33% evidencing

increases. Taken together, these studies suggest ASD symptom stability rates of 50-70%

within FXS. However, inferences about the diagnostic stability of clinical diagnoses of

ASD within FXS are limited due to the reliance on the use of measure cutoffs to create

diagnostic groups, rather than clinical decision making, in most current published work.

A more recent series of studies has emerged examining the stability of autism

features in infants with FXS. One of these studies used a case study approach to compare

8 participants between 9-24 months of age with FXS+ASD and FXS-noASD outcomes at

24 months of age (Hogan et al., 2017). Although the sample size was small, comparisons

suggested that autism symptomatology as measured by the Autism Observation Scale for

Infants (AOSI, Bryson, Zwaigenbaum, McDermott, Rombough, & Brian, 2008)

decreased in all participants between 9 and 12 months of age, with lower mean scores for

the participants in the FXS-noASD group. In a complementary study, similar patterns

were found in consistency of autism symptomatology between the 12 and 24 month time

points, in that infants with ADOS scores above diagnostic cutoffs at 24 months evidenced

higher AOSI scores at 12 months of age, than those who did not meet ADOS diagnostic

cutoffs at 24 months (Roberts, Tonnsen, McCary, Caravella, & Shinkareva, 2016). These

papers are important additions to the literature, in that they are the first to begin to

8

explore autism symptomatology and stability in infants with FXS syndrome, when autism

symptoms often first emerge.

Taken together, these studies suggest variability in autism symptomatology across

development in young males with FXS, leading to inconsistent reports of diagnostic

stability. This variability may reflect the dynamic nature of development suggesting that

symptoms of ASD vary across time in individuals with FXS as other development

processes are occurring, such as increases in language/communication abilities or

increasing social demands. However, the significant variability in measurement tools and

diagnostic practices (i.e. measure cutoffs vs. clinician judgment) may also be contributing

to the inconsistent findings. Lastly, change in autism symptomatology in individuals with

FXS may be subtle throughout development and therefore poorly captured by measures

designed to produce categorical outcomes.

Challenges measuring change in autism symptomatology over time is not a

unique problem in FXS research (Anagnostou et al., 2015). Studies examining change in

core autism symptomatology in nsASD are also limited given that the gold standard

measures used in many studies such as the ADOS or ADI-R are diagnostic measures,

with limited score ranges that are intended to measure presence or absence of ASD

symptoms. A new measure, the Brief Observation of Social Communication Change

(BOSCC, Grzadzinski et al., 2016) has been designed specifically to address these

limitations in the ability to detect subtle change in autism symptomatology over time in

current commonly used measures. The BOSCC measures many of the core features

identified on the ADOS and ADI-R, however the coding ranges allow for greater

variability, allowing for detections of subtler changes. Currently, two studies using the

9

BOSCC have been published; a paper examining the psychometric properties of the

BOSCC (Grzadzinski et al., 2016) and a paper examining ASD symptom change

following an early intervention program in 21 children with nsASD (Kitzerow, Teufel,

Wilker, & Freitag, 2016). In both studies, changes in BOSCC scores were detected with

small to medium effect sizes in as little as 6 months to 1 year, suggesting that the measure

is robust to detecting findings over short periods of time, and in small samples (n = 21;

Kitzerow et al., 2016).

Current Limitations

Across this body of work significant inconsistencies are present in the measures

used to quantify change in autism symptomatology, which make comparisons across

studies challenging. This is especially difficult given that some are parent report

measures while others are based on trained clinician observations. Additionally, all

measures reported (e.g. CARS-2, ADOS, ADI-R, AOSI), were not developed to measure

change, making clinically significant change harder to identify. For example, Hatton and

colleagues (Hatton et al., 2006) finding that CARS-2 scores increased 2 points over 10

years may have limited functional significance. Given that ASD is theoretically a lifelong

condition, this small change in total score over a 10-year period may instead reflect

reliability.

Other commonly used measures, such as the ADOS and ADI, are diagnostic tools

that are not sensitive enough to capture change over short periods of time (Anagnostou et

al., 2015; Grzadzinski et al., 2016). Given the diagnostic intention of these tools, some

items are considered to be “low threshold”, in that if the symptom is present, it is coded.

This approach, similar to a “present” or “absence” rating, it is not designed to

10

differentiate nuances in autism symptomatology that may be expected to change

gradually over time. This is in contrast to the design of the BOSCC, which includes a

greater scoring range (0-5 vs. 0-3) and the ability to capture subtleties in the frequency

and quality of social communication behaviors and restricted interests/repetitive

behaviors. Additionally, the commonly used practice of using the ADOS or ADI to both

create the groups of interest (i.e., FXS+ASD vs FXS-noASD) and measure differences

(i.e., how do these groups differ on ASD symptoms as measured by the ADOS?), is

inherently flawed in the circular nature of these questions (i.e., if the ADOS was used to

differentiate the groups of interest, we would expect the groups to be different on the

ADOS). Therefore, a tool such as the BOSCC, which measures related but not identical

constructs to gold standard diagnostic tools of ASD, reduces the consequences of the

previously described practice.

The Present Study

The present study aims to address these gaps by utilizing a new measure designed

to capture change in autism symptomatology in a sample of 2 to 6-year-old males with

FXS. The BOSCC has been designed specifically to address limitations of current

measures’ ability to accurately measure change in core features of autism. Additionally,

predictors of these trajectories will be examined to determine both risk and protective

factors of the development of impairing autism symptomatology within FXS. By looking

at a range of predictors that have been previously identified as being related to co-morbid

autism symptomatology in FXS, we aim to identify risk factors that may be important

targets for intervention to improve functioning in young children with FXS. Lastly, this

11

study aims to add to the literature by examining trajectories based on diagnostic stability

patterns, rather than using a single diagnostic timepoint to determine grouping status.

The present study aims to address the following research questions:

1) Do patterns of change in autism symptomatology differ in young males with FXS

dependent on their CBE diagnosis of ASD, and the stability of that diagnosis?

a. We hypothesize that children with FXS who have a CBE diagnosis of

ASD will show increasing ASD symptom severity over time, while

participants with FXS who do not have ASD will show minimal change in

the severity of ASD symptomology.

2) Do cognitive and language functioning in males with FXS predict change in

autism symptomatology over time?

a. Consistent with previous literature, we hypothesize that lower cognitive

and language functioning will be predictive of increasing autism symptom

severity.

Methods

Participants

This study included 28 males with FXS assessed between the ages of 2-8 years.

Females were excluded from the analysis because of small sample available for analysis

(n=11), with only 1 female meeting diagnostic criteria for ASD. Each participant had at

least 2 diagnostic assessments using the ADOS-2. The participants are drawn from a

larger ongoing longitudinal study, herein referred to as the “parent study”, measuring the

developmental trajectories of infants and preschool age children with FXS. Inclusion

criteria for the study included gestational age >36 weeks, English as the primary language

12

in the home and no other known medical conditions. Participants were recruited primarily

from research and medical sites, as well as social media sites, specializing in FXS.

The initial parent study measured the development of autism symptomatology in

infants with FXS (R01MH90194, PI: Roberts). Participants in this study were seen at 9,

12 and 24 months with a diagnostic assessment for ASD completed at the 24-month time

point. The subsequent renewal of this study allowed for the continued longitudinal

examination of the development of ASD in these participants through the preschool years

with diagnostic assessments for ASD targeted annually at 3, 4 and 5 years-of-age.

Therefore, the majority of participants had their first ASD diagnostic assessment at 24

months-old and follow up assessments at 3, 4 or 5 years-of-age. There was variability in

the number of assessments and in the age at which assessments occurred, however, given

the age of participants at onset of the study renewal and inclusion of a handful of

participants who were recruited for enrollment starting at 4 or 5 years-of-age. For

inclusion in the present study, participants were required to have at least 2 video recorded

diagnostic assessments which included the Autism Diagnostic Observation Schedule-2.

Measures

Brief Observation of Social Communication Change. The Brief Observation of

Social Communication Change (BOSCC, Lord et al., 2016) is a coding scheme designed

to measure change in autism symptomatology over time, when compared across 2 or

more time points. The BOSCC is coded based on 10-12 minutes of semi-structured play

between an adult and the target child. The BOSCC can be applied to a range of play

interactions that meet basic criteria by providing engaging and developmentally

appropriate materials and allowing the child to move freely around the room and explore

13

toys. Additionally, whatever setting and structure is selected for the initial observation

should be kept consistent for all future play sessions to which the BOSCC will be coded

and compared. The coding scheme consists of item level decisions trees that allow the

coder to answer a series of yes or no questions based on the behaviors in the observation.

Observations are split into two 6-minute segments, which are coded independently.

Fifteen items are coded, and items 1-12 are included in the total score. Total scores across

the two segments are then averaged to determine a participant’s total score on the

BOSCC. For the present study, the BOSCC coding scheme will be applied to video clips

selected from the participant’s ADOS-2 administrations. Video clips include the

following tasks from the ADOS-2 administrations; 3 minutes of Free Play, 3 minutes of

Bubble Play, 3 Minutes of Birthday Party or Bath Time and 3 Minutes of Balloon Play.

Alternative tasks are available to supplement (e.g. Snack, Response to Joint Attention) if

these tasks are less than 3 minutes in length. This collection of tasks and time lengths is

recommended by the authors and has been previously published by Kitzerow and

colleagues (Kitzerow et al., 2016). The first author of the present study has attended

research training on the coding of the BOSCC and has been given authorization to use it

for research purposes by the authors. According to measure use procedures, a team of

undergraduate research assistants were trained up to reliability standards with the first

author. To achieve reliability with the first author, coders must be within 1 point for more

than 80% of items on each coding segment and within 4 points on each segment total

score. To minimize coding drift, 20% of videos were coded by the first author and

another coder. Intraclass correlations were used to examine reliability of absolute

agreement, which was high (ICC = .911). BOSCC scores collected from the same child

14

over time are compared to measure change. BOSCC scores will be the primary outcome

variable in the analyses.

Autism Diagnostic Observation Schedule-2. The Autism Diagnostic

Observation Schedule-2 (ADOS-2) is a standardized semi-structured play-based measure

used to assess behaviors consistent with a diagnosis of ASD. To administer the ADOS-2

within a research setting, examiners are required to attend research training and achieve

reliability in the administration and coding of the measure with group of certified trainers

at 80% reliability or become reliable with a research reliable examiner at the research

site. The ADOS-2 has 5 modules, which are administered based on an individual’s

chronological age and expressive language level. Throughout the ADOS-2, the examiner

provides opportunities for social interaction through a series of standardized behavioral

presses and activities. Some of these activities include free play, bubble play, balloon

play and imaginative play with a baby doll. Behaviors observed during these presses are

then coded on items measuring core features of ASD. Items are generally scored on a

scale of 0-3, with scores of 0 indicating a normative response or the absence of

atypicality, and a 3 indicating severe atypicality indicative of ASD. Items identified as

being the most specific to an autism diagnosis from each module are used to create an

algorithm score. A calibrated severity score can be derived from the algorithm score to

allow comparison across the ADOS modules. Participant scores on the ADOS-2 are used

to inform the CBE diagnostic procedures to determine the presence or absence of autism

at each visit. The ADOS-2 reports strong inter-rater reliability of 84% (Lord et al., 2012).

Within lab inter-rater reliability is calculated on a randomly selected sample of 20% of

administrations as part of the completed and ongoing studies. Current within lab inter-

15

rate reliability is 83.3%. As previously mentioned, segments of the videotaped ADOS-2

administrations will be used for the BOSCC coding scheme.

Mullen Scales of Early Learning. The Mullen Scales of Early Learning

(Mullen, 1995) is a standardized measure of development normed for young children

ages 0-60 month, that measures development in five areas; Gross Motor, Fine Motor,

Receptive Language, Expressive Language, and Visual Reception. Each domain produces

a raw score, which can be transformed into a T-Score and an Age Equivalent Score. T-

Scores have a mean of 50 and a standard deviation of 10. A summary standard score, the

early learning composite (ELC), represents a composite score across all domains, with

the exclusion of Gross Motor. The ELC has a mean of 100, and standard deviation of 15.

Limitations exist in the use of T-scores for young children with developmental

disabilities due to floor effects. In this sample, in the domains of Visual Reception,

Expressive Language and Receptive Language, 46%, 54% and 64% were at the floor for

each domain respectively. Therefore, Mullen raw scores will be used in the analysis as

predictors of change scores on the BOSCC. The sum of the scores on the Expressive and

Receptive language domains will be used as a measure of language functioning. The

Visual Reception domain will be used as a measure cognitive functioning. Standard

scores will be used for describing the sample. Internal consistency for each of the skills

ranges from 0.75 to .08, and coefficients of test-retest reliability range from .70- .80

(Mullen, 1995).

Procedure

The USC institutional review board has approved all data collection procedures.

Assessments were conducted primarily in the participant’s homes, with occasional visits

16

at the Neurodevelopmental Disorders Lab at USC based on family preference.

Conducting visits in the participants’ homes allowed for decreased travel burden on

families with children with disabilities and supports increased ecological validity as it

removes the potential confound associated with a child’s adjustment to a novel laboratory

setting. Procedures followed in the home and lab settings were identical. Assessments

were conducted over the course of one to two days, in 3-4-hour sessions each day,

depending on the child’s age and attention span. Breaks were given throughout the

assessment as needed. Participants were compensated for participation. All in-person

measures were administered by research staff that included bachelors level research

assistants, graduate students and post-doctoral researchers. ADOSs were only

administered by staff who had achieved independent research reliability or within lab

reliability with a research reliable counterpart. For all assessments included in the present

study, a CBE diagnosis of ASD was determined at each time point by a team of at least 3

members of the research staff who were all research reliable on the ADOS, one of whom

was a licensed psychologist. The CBE procedure is standard in the field of ASD research

(Ozonoff et al., 2015). Data from the Mullen, Vineland Adaptive Behavior Scales, and

ADOS was integrated to determine whether a child met criteria for ASD. Additionally,

all research reliable staff on the ADOS attended monthly consensus coding meetings for

ongoing calibration of scoring.

Clinical best estimates of ASD were determined for participants at each time point

and the ADOS-2 videotapes were scored using the BOSCC. In addition to the evaluation

of ASD, global developmental delay was evaluated for each participant. All participants

in the study met criteria for global developmental delay at each assessment which is

17

consistent with the phenotype of young males with FXS. Therefore, all participants in

this study who meet criteria for ASD also have co-morbid developmental delay at all time

points.

Data Analytic Plan

At the time of analysis, 28 males with FXS had 2 or more ADOS administrations

available for BOSCC coding and were included in the analysis (total of 73 assessments,

range of 2-4 assessments per participant). Multilevel models are robust to variability in

the number of assessments. As an initial step, data were plotted longitudinally using

spaghetti plots. A linear pattern was evident, therefore linear models are used. Multilevel

models were used to examine change over time in BOSCC scores, specifying time as the

level 1 clustering variable, with individual variables entered at level 2. Across all models,

age was centered at 24 months, given that the majority of participants had their first time

point at 24 months. Therefore, intercepts are interpreted as the average BOSCC score at

24 months of age. Slopes will be interpreted as the average change in BOSCC scores per

month.

As an initial step, participants were grouped in terms of the stability of ASD

diagnoses. A stable diagnosis was defined as meeting or not meeting ASD diagnostic

criteria at all assessments while an unstable diagnosis was defined as meeting ASD

diagnostic criteria for one or more assessments but not across all assessments. In the total

sample, 54% percent had stable ASD positive diagnoses, 21% percent had stable ASD

negative diagnoses, 11% percent changed from an ASD positive diagnosis to an ASD

negative diagnosis, and 14% percent changed from an ASD negative diagnosis to an ASD

positive diagnosis. Thus, 75% demonstrated a stable positive or negative ASD diagnosis

18

and 25% demonstrated an unstable ASD diagnostic trajectory representing both initial

false positives and false negatives. Based on these diagnostic stability patterns,

participants were split into 4 groups: 1) stable ASD positive (Stable ASD; n= 15), 2)

stable ASD negative (Stable noASD; n= 6), 3) participants with an initial diagnosis of

noASD which changed to ASD (Unstable ASD; n=4), and 4) participants with an initial

diagnosis of ASD which changed to noASD (Unstable noASD; n =3). Fit statistics (e.g.,

AIC and BIC) for models using the diagnostic stability groups described previously were

compared to models that dichotomized diagnosis to ASD or noASD based on the

participants “outcome” diagnosis (i.e., their diagnosis at the last time point for the

participant in the study). The model using the diagnostic stability grouping structure

evidenced the best statistical fit and aligned with the theoretical interest of the study, and

previous work in nsASD where diagnostic stability has been examined (Ozonoff et al.,

2015). Therefore, to answer research question one, the stability diagnostic group was

entered into the model as a level 2-time invariant predictor, to determine how diagnostic

group impacts initial BOSCC scores at 24 months and change over time.

To answer research question two, raw scores on the Mullen were entered in the

model as level 2-time invariant predictors. Raw scores were selected from the

participants assessment closest to the age of 24 months (m = 32.5, range = 23-78

months). Raw scores for expressive and receptive language were highly correlated (r =

0.79), therefore they were combined (i.e., summed) to create one language variable. A

sum, rather than an average, was selected to create the composite language variable given

that scores across the two language domains do not share an equivalent scale (i.e., 6 raw

score points on the Expressive Language domain is not necessarily developmentally

19

equivalent to 6 raw score points on the Receptive Language domain). Visual Reception

raw scores were used as a measure of early non-verbal cognitive functioning.

Results First, ANOVA tests were run to determine if the groups differed on demographic

variables or predictor variables of interest. All means and standard deviations are

presented in Table 1. More participants in the Stable ASD and Unstable noASD groups

were older than 24 months at their first ADOS/BOSCC time point (2/3 in the Unstable

noASD group, and 4/15 in the Stable ASD group), however differences in mean ages

between groups were not found to be statistically significant. Additionally, no statistically

significant differences were found between groups on age at Mullen, Receptive Language

T-scores or Fine Motor T-scores. For the remaining domains, overall F-tests were

significant. Therefore, Tukey HSD post-hoc tests were run to determine where

differences lie between groups. Only significant post-hoc tests are reported. On the

Expressive Language domain, the Unstable ASD group scored 11 points higher than the

Stable ASD group (p = .015). On the Visual Reception domain, the Stable noASD group

scored 8 points higher than the Stable ASD group (p = .053). The Early Learning

Composite was lowest in the Stable ASD group, which was 9.31 points lower than the

Stable NoASD group (p = .000), 8.79 points lower than the Unstable ASD group (p =

.00), and 12.11 points lower that the Unstable noASD group (p = .000). Participants in

the Stable ASD group had an average ADOS calibrated severity score (CSS) of 8.6,

which was 6 points higher than the Stable NoASD group (p = .000), and 4 points higher

than the Unstable ASD group (p = .000). Additionally, the Unstable NoASD group had

20

an average CSS score of 7.67, which was 5 points higher than the Stable NoASD group

(p = .000), and 3 points higher than the Unstable ASD group (p = .015).

A taxonomy approach to the multilevel modeling was taken to address the

proposed research questions (Singer, Willett, & Willett, 2003). Results of the series of

models are presented in Table 2. First, the unconditional means model and unconditional

growth model were run. The unconditional means model (Model A) produced an

intraclass correlation (ICC) of 67.33, suggesting that 67% of the variance in BOSCC

scores are attributable to differences among participants. Next, the unconditional growth

model, including age centered at 24 months, was run. Within-person variance declined

between Model A and Model B, producing a pseudo R2 of 0.38, suggesting that 38% of

the within-person variability is associated with the linear effect of time.

Next, diagnostic group was added to the model as a predictor of initial status (i.e.,

average BOSCC score at 24 months) and change over time (Model C). The Stable ASD

group was selected as the reference group. On average, participants in the Stable ASD

group had initial BOSCC scores of 31.64 (t = 14.40, p = 0.00). Participants in the Stable

NoASD group (b = -16.49, t = -4.20, p = 0.00), and Unstable ASD group (b = -9.86, t = -

2.24, p = .034) both had lower initial BOSCC scores than the Stable ASD group. Initial

BOSCC scores did not differ between the Unstable NoASD group and the Stable ASD

group (b = 0.68, t = 0.10, p = .924). Slopes did not differ between the Stable ASD group

and the Stable NoASD or Unstable ASD groups. However, participants in the Unstable

NoASD group evidenced a negative slope (b = 0.42, t = -1.90, p = .065), suggesting that

their BOSCC scores decreased by 0.42 points per month, compared to the Stable ASD

group. These trajectories are displayed in Figure 1. Between person variance declined

21

from Model B to Model C producing a pseudo R2 of 0.60, which suggests that 60% of the

variability in initial BOSCC scores is associated with diagnostic group.

To answer research question 2, additional models were run to examine the effects

of language and non-verbal cognitive ability on initial BOSCC scores and change over

time. Language was determined to be negatively related to initial BOSCC scores (b =-

0.49, t = -3.93, p = .001), but not change over time. Non-verbal cognitive ability was not

related to initial BOSCC scores or change over time. When non-verbal cognitive ability

was added to the model, language remained as a significant predictor, however model fit

declined (i.e., AIC and BIC estimates increased), so non-verbal ability was removed for

the final model. Variance estimates for Model D suggest that language accounts for an

additional 8% of the variability in initial BOSCC scores.

Discussion

This study examined the development of autism symptomatology over time in

young males with FXS, using a novel measure of social communication change.

Multilevel models were used to examine linear change based on diagnostic group,

including early language and non-verbal cognitive ability as predictors of initial BOSCC

scores and change over time. The data included in this study are drawn from a parent

study that is the first study to track the stability of the clinical best estimate diagnosis of

ASD in young children with FXS, beginning in toddlerhood (NIH R01MH90194, PI:

Roberts). Therefore, this study is the first to explore the longitudinal development of

autism symptomatology change in toddler and preschool aged children with FXS

accounting for instability in the diagnoses of ASD at the individual level. While limits

exist with this approach (i.e., namely, small sample sizes), the information gathered from

22

this study is essential for furthering the understanding of the heterogeneity of autism

Participants comprised 4 groups, Stable ASD (54%), Stable NoASD (21%),

Unstable ASD (14%) and Unstable NoASD (11%). For those who received their first

diagnosis at 24 months (n=22), 77% had stable diagnostic trajectories and maintained the

diagnosis of either ASD or non-ASD across time. Of the 5 participants who displayed

ASD diagnostic instability, 1 was a false positive, and 4 were false negative. This

suggests that it is more likely for young males with FXS to be false negative than false

positive for ASD at 2 years old. This is consistent with patterns in nsASD work of a 5:1

false negative to false positive ratio (Ozonoff et al., 2015). At 24 months, 83% of

participants who received a diagnosis of ASD maintained that diagnosis across time

which is consistent with the nsASD literature, which reports a diagnostic stability of ASD

at 24 months of 82% (Ozonoff et al., 2015).

Our results suggest that there is significant variability in the presence and severity

of ASD symptomatology within the FXS phenotype at 24 months-of-age. This

variability can be partially explained by diagnostic categorization and stability of ASD.

As a group, young children with FXS who have early and stable diagnoses of ASD

throughout childhood display the most severe symptomatology, with symptom levels

remaining high and stable across time. In a seemingly parallel pattern, participants with

an early and stable diagnosis of FXS noASD, evidence much lower levels of autism

symptomatology at 24 months with symptoms remaining low and stable over time. This

finding is contradictory to the results published by Hernandes and colleagues (2009), who

reported that autism symptomatology across their two diagnostic groups (FXS+ASD and

FXS-noASD) became less differentiated across time. These contrasting findings may be

23

attributable to differing methodologies across the studies in the measurement of autism

symptoms (i.e., BOSCC in the present study vs. ADI-R in the Hernandes study),

diagnostic categorization procedures (i.e., CBE in the present study vs. ADI-R cutoffs at

Time 1 in the Hernandes study), or their slightly older sample at Time 1 (i.e., 33 months

in the present study vs. 56 months in the Hernandes study). Additionally, Hernandes and

colleagues reported that 32% of participants in their sample changed diagnostic

categories over the course of the study, however their diagnostic grouping was based on

their Time 1 data. Therefore, their dichotomous approach subsumed participants with

instability in their diagnostic trajectories into the diagnostic group determined at their

initial visit. The findings presented in the current study may represent the trajectories for

more “pure” diagnostic groups of FXS+ASD and FXS-noASD, given that the Stable

ASD and Stable NoASD groups exclude participants who display diagnostic instability.

Compared to the Stable ASD group, participants in the Unstable ASD group had

initial BOSCC scores that were 10 points lower, however there were no measurable

differences in slopes. In theory, the UnstableASD group represents a group of males with

FXS whose early symptom presentation did not warrant a diagnosis of ASD, with

symptoms appearing more consistent with ASD over time (i.e., false negative cases).

Contrary to what was hypothesized, these participants’ BOSCC scores didn’t

significantly “worsen” with time, rather they follow a flat trajectory. Clinically, this

group may represent a group of males with FXS who exhibit a milder presentation of

ASD and whose symptoms of ASD are harder to distinguish from global developmental

delay in toddlerhood. The flat BOSCC trajectories would suggest that their diagnostic

instability may be influenced by clinician factors (e.g., a more conservative approach to

24

diagnosing ASD in FXS in toddlerhood when the symptom presentation is less severe

and global developmental delay is present), rather than a result of a significant

“worsening” in symptom presentation with time.

This finding is distinct from what has been described in the nsASD literature,

where the “false negative” participants evidenced a worsening of autism symptomatology

over time. This worsening is hypothesized to represent the later emergence of ASD,

rather than clinician level variables or true misdiagnoses at earlier timepoints (Ozonoff et

al., 2015). While autism symptom trajectories appear to differ, the false negative groups

across the current study and the nsASD literature share a common feature of being

slightly higher functioning in language and cognitive abilities than those with early and

stable diagnoses of ASD.

The only group to display a declining slope (i.e., fewer ASD symptoms over time)

was the Unstable NoASD group. These participants are similar to participants described

as “false positives” in other work, who receive a diagnosis of ASD early on in

development, however they no longer meet criteria for ASD as they age. This group was

the minority of the sample, with only 3 participants falling into this group. Clinically,

these participants are important as they represent a group of children who appear to “get

better” with time, who may hold answers to effective interventions or protective factors

within the FXS phenotype. In contrast to false positive cases in the nsASD literature,

these children do not show a milder or “intermediate” presentation of ASD. Initial ADOS

severity scores for the Unstable NoASD group were 7.67, which is on the cusp between

moderate to high levels of ASD related symptoms and did not differ from the Stable ASD

group. However, these two groups did differ on global development, with the Unstable

25

NoASD group scoring almost 1 standard deviation higher on a measure of broad

development at their initial time point, measured by the Mullen Early Learning

Composite.

In addition to understanding developmental trajectories of autism

symptomatology, this study aimed to discover predictors of early symptom presentation

and trajectory, with non-verbal cognitive functioning and language being the primary

domains of interest. Language and non-verbal cognitive impairment have long been

established as prominent features of the FXS phenotype, often developing at similar

levels (Abbeduto, Brady, & Kover, 2007). These delays complicate the diagnosis of

ASD in FXS given the need to differentiate global delays in development from a specific

social communication impairment consistent with ASD. For example, a clinician may

have trouble determining whether a lack of gesture use or poor integration of

communication forms (e.g., eye-contact, gestures, vocalizations) is the result of an

immature social communication system or an impaired one resulting from ASD. Given

this concern, clinicians may attribute autism symptomatology seen within FXS to

developmental delay, more broadly.

In this study, higher language scores at 24 months, predicted lower initial BOSCC

scores (i.e. less autism symptomatology), but not slopes over time. In contrast, non-verbal

cognitive ability was not predictive of initial BOSCC scores or slopes over time.

Therefore, early language abilities, but not non-verbal functioning, may serve as a

protective factor for children with FXS resulting in more robust and advanced social

communication abilities. Language abilities have also been found to predict ASD

symptom severity in adolescents with FXS, suggesting that this is a persistent

26

relationship within the FXS+ASD phenotype (Abbeduto et al., 2019). This finding has

important implications for early intervention targets in young children with FXS, with

recommendations to focus on language development before 24 months as a possible

protective factor to the development of autism symptomatology.

Given that scoring on two BOSCC items includes vocalizations (i.e. vocalizations

directed to others, integration of vocal and non-vocal modes of communication), it is

prudent to consider whether the relationship between higher language abilities and

BOSCC scores may be influenced by these items. The current sample is not large enough

to conduct this type of analysis, however future research may consider examining if

overlap in the measurement of language on the Mullen and BOSCC is able to partially

explain this relationship. However, vocalizations need not be complex to receive credit

on the BOSCC, therefore it is unlikely that the relationship between language abilities

and initial BOSCC scores can be completely accounted for by these items.

While the variables of interest in this study accounted for a significant amount of

variance in this data, unexplained variance remains to be understood. Specifically, we

were unable to explain some of the variance associated with change over time in autism

symptomatology. Future work may choose to focus on time-varying predictors which

may be able to provide additional insight into predictors of individual growth trajectories.

This study presents important and novel findings to enhance the understanding of

autism symptom trajectories within the early developmental period in males with FXS.

Individuals with FXS and co-morbid ASD are at greater risk for more profound

impairment in language, cognition and adaptive skills, and require greater levels of

service to achieve optimal outcomes. The findings presented here inform not only

27

possible surveillance and diagnostic practice of ASD in FXS, but also inform early

intervention recommendations. The vast majority (83%) of participants who received a

diagnosis of ASD at age 2, maintained this diagnosis over time. This finding supports the

clinical practice of screening for, evaluating, and diagnosing ASD within FXS as early as

age 2, should symptoms be clear and compelling. Additionally, clinicians should remain

vigilant and follow children whose presentation at 24 months is atypical, albeit milder

than FXS peers with more impaired cognitive and language abilities. This study

highlights that these individuals may not necessarily “get worse” over time or grow into a

diagnosis of ASD. Rather, a lack of improvement in social communication or a lack of

any clear reduction in autism symptomatology may be evidence of co-morbid ASD. It is

important to note that the clinicians in this study were highly experienced with both ASD

and FXS, and this level of expertise in unlikely to be present in all community settings

where families of young children with FXS may be seeking an evaluation for ASD. This

may increase the risk for diagnostic overshadowing in these settings, where clinicians

may be less likely to diagnose a co-morbid condition instead attributing behavioral

symptoms of ASD to the FXS phenotype.

Early language abilities, distinct from non-verbal abilities, were predictive of

BOSCC scores at 24 months with higher language scores being associated with lower

BOSCC scores. Language abilities at 24 months, however, did not make a child more or

less likely to have a declining or inclining slope. Given the longstanding knowledge of

language impairment in FXS, and the relationship identified between language and early

autism symptomatology in this study, language intervention should be prioritized as an

early intervention for young children with FXS.

28

This study is limited by a few factors. Primarily, the small sample size. Fragile X

syndrome is a rare syndrome, making recruitment of large samples especially at young

ages, challenging. This sample capitalizes on longitudinal data from 28 males which

results in 73 data points providing more power to the analysis. Admittedly, the diagnostic

grouping structure (i.e., stability groups) results in very small samples for comparison.

However, this structure aligned with the theoretical questions of this study and resulted in

a better statistical fit than a dichotomous approach of FXS+ASD and FXS-noASD.

Future work should assess generalizability of the findings in a larger sample. Females

were intentionally excluded from this study given the significant heterogeneity within the

female FXS phenotype. Therefore, we must acknowledge that the results presented here

can only be generalized to young males with FXS.

Table 1.1 Participant Demographics for FXS Males (n = 28)

Stable ASD

n = 15

Stable noASD

n = 6

Unstable

ASD

n = 4

Unstable noASD

n = 3

ADOS-2/BOSCC1

Age in Months 35.15 (18.08) 25.63 (0.75) 24.77 (1.35) 43.32 (16.06)

ADOS CSS 8.6 (1.4) 2.33 (1.21) 3.75 (2.99) 7.67 (0.58)

BOSCC Total 30.63 (9.41) 15.08 (6.76) 20.25 (1.04) 25.50 (11.79)

Mullen2

Age in Months 29.15 (11.35) 25.65 (0.70) 24.57 (1.35) 39.21 (12.90)

ELC 52.33 (6.62) 60.83 (7.96) 59.75 (3.20) 63.33 (13.58)

Visual Reception T-Score 22.87 (5.11) 31.00 (7.21) 31.00 (2.58) 33.33 (11.55)

Exp. Language T-Score 22.53 (5.37) 28.00 (7.18) 34.00 (3.37) 30.33 (10.50)

Rec. Language T-Score 22.53 (6.52) 29.50 (12.71) 22.50 (3.00) 29.00 (8.54)

Fine Motor T-Score 21.47 (3.23) 24.33 (4.59) 22.75 (1.89) 25.00 (5.57)

ADOS-2 Autism Diagnostic Observation Schedule, 2nd Edition, BOSCC Brief Observation of Social Communication Change, CSS Calibrated Severity Score, Mullen Mullen Scales of Early Learning, ELC Early Learning Composite 1Data corresponds to the first ADOS available for coding 2Data corresponds to the Mullen administered closest to the 24-month timepoint; standard scores are presented.

29

Table 1.2 Results of a taxonomy of multilevel models for change in BOSCC scores over time Model A Model B Model C Model D Initial Status Intercept 25.80**

(1.94) 25.06** (2.16) 31.64** (2.20) 41.32** (3.07)

Stable noASD -- -- -16.49** (3.92) -10.22** (4.07) Unstable ASD -- -- -9.86** (4.40) -4.80 (4.33) Unstable

noASD -- -- -0.68 (7.12) 11.45 (7.47)

Language -- -- -- -0.49** (0.13) Rate of Change Intercept -- 0.04 (0.06) 0.07 (0.08) 0.12 (0.08) Stable noASD -- -- -0.21 (0.23) -0.33 (0.22) Unstable ASD -- -- -0.00 (0.16) -0.04 (0.16) Unstable

noASD -- -- -0.42* (0.23) -0.41* (0.22)

Variance Components

Level 1 Within person 42.62 (6.52) 26.48 (5.14) 26.63 (5.16) 26.05 (5.10) Level 2 (b/t person)

In initial status 87.86 (9.37) 85.74 (9.25) 34.08 (5.84) 27.73 (5.27)

In rate of change

-- 0.04 (0.21) 0.05 (0.22) 0.05 (0.22)

Pseudo R2 Within person -- 0.38 0.38 0.38 In initial status -- -- 0.60 0.68 Model Fit AIC 534.07 536.62 512.71 504.96 BIC 540.90 550.36 538.80 533.03 ** p < .05, *p < .1

30

31

Stable ASD

Stable noASD

Unstable ASD

Unstable noASD

Figure 1.1 Uncontrolled effects of diagnostic stability group on BOSCC scores over

time

32

CHAPTER 2

CROSS-SYNDROME CONTRASTS OF AUTISM SYMPTOMS IN

YOUNG CHILDREN: FRAGILE X, DOWN SYNDROME AND NON-

SYNDROMIC AUTISM

Autism spectrum disorder (ASD) is a developmental disorder that is characterized

by deficits in social communication and the presence of atypical restricted and repetitive

behaviors and interests (American Psychiatric Association, 2013). ASD is one of the

most common and impairing childhood conditions, with prevalence rates of about 1.5%

in the general population (Centers for Disease Control, 2014). Heritability of ASD is

estimated to be around 70-80% with more than 1000 identified risk genes as well as a

handful of single gene disorders that carry a high risk for developing ASD (Geschwind,

2011).

ASD is a heterogeneous disorder, with core impairments of social communication

and restricted and repetitive behaviors accompanied by a myriad of cognitive, language

and adaptive behavior presentations. One of the most impairing co-morbid features of

ASD is intellectual disability (ID), which occurs in approximately 37-55% of individuals

with ASD (Charman et al., 2017; Rivard, Terroux, Mercier, & Parent-Boursier, 2015).

The presence of ID in young children at risk for ASD can make diagnosing ASD more

complex, given that it requires diagnosticians to differentiate delays associated with ID

from those better explained by ASD (Matson & Shoemaker, 2009). Diagnostic

differentiations are also complicated by the overlap in symptomatology between ASD

33

and ID including symptoms often thought to be classic features of ASD such as repetitive

behaviors and echolalia occurring commonly in individuals with ID more broadly

(Wallander, Dekker, & Koot, 2003). However, accurate and early diagnosis of the co-

morbidity of ASD and ID is essential, given that their co-occurrence causes individuals to

experience greater impairments and less optimal outcomes than individuals with either

condition in isolation.

Single-Gene Models of ASD

Single gene disorders account for about 1-2% of all ASD cases (Abrahams &

Geschwind, 2008). While mechanisms for impairment across the single-gene disorders

are varied (e.g. excess gene expression, protein deficit) consistency exists in the almost

universal ID present in many conditions. Therefore, single gene disorders are an excellent

model for studying the behavioral phenotype of ASD in individuals with ID. Fragile X

syndrome (FXS) and Down syndrome (DS) are two excellent models to examine these

relationships given that ID is nearly universal in both disorders, and they are relatively

common in the population at 1 in 5000 (Coffee et al., 2009) and 1 in 1500 (Kazemi,

Salehi, & Kheirollahi, 2016) respectively.

Autism Symptomatology in FXS

FXS is the most common form of inherited ID (Hagerman & Hagerman, 2002).

FXS results from a repeat expansion of the CGG sequence on the FMR1 gene on the X

chromosome. Repeats on the FMR1 gene that exceed 200 result in methylation of the

gene, which significantly reduces or eliminates production of Fragile X Mental

Retardation Protein (FMRP), which is essential for cognitive development. Females with

FXS are often less impaired than males with FXS due to the presence of a second X

34

chromosome which will produce FMRP if it is unaffected by the FXS mutation

(Hagerman & Hagerman, 2002).

A significant body of work has been developed examining the relationship

between ASD and FXS. Consistent across these studies are findings that the rates of ASD

in FXS are high, with estimates ranging from 60-75% (Harris et al., 2008; Klusek,

Martin, & Losh, 2014). The presence of ASD in FXS results in greater impairment across

a range of domains including cognition (Bailey Jr, Hatton, Tassone, Skinner, & Taylor,

2001; Hogan et al., 2017; Kaufmann et al., 2004), adaptive behavior (Caravella &

Roberts, 2017; Hahn, Brady, Warren, & Fleming, 2015), language (Klusek et al., 2014),

motor skills (Bailey, Hatton, Mesibov, Ament, & Skinner, 2000) and social approach

behaviors (Roberts et al., 2019). There is consensus that features consistent with ASD are

commonly observed in individuals with FXS (Hagerman & Hagerman, 2002), however

there continues to be debate about the etiology of the symptoms, and whether or not they

constitute the presence of a true comorbidity of ASD or represent phenotypic features

associated with FXS. Evidence across biological and behavioral studies suggests that

symptoms of ASD in FXS may be behaviorally similar to non-syndromic ASD (nsASD),

however different patterns of strengths and weaknesses may emerge with time (Wolff et

al., 2012) that could be driven by unique brain mechanisms (Hazlett, Poe, Lightbody,

Gerig, Macfall, et al., 2009). These strengths in FXS and comorbid ASD (FXS+ASD) are

typically identified in the social domain, marked by increased frequency and quality of

social overtures even in those diagnosed with ASD (McDuffie, Thurman, Hagerman, &

Abbeduto, 2015; Wolff et al., 2012). However, evidence suggests that strengths in social

communication in FXS+ASD may not emerge until around the age of 4 of 5 (Wolff et al.,

35

2012), with little differentiation between nsASD and FXS+ASD observed in the toddler

years (Hazlett, Poe, Lightbody, Gerig, Macfall, et al., 2009; Rogers, Wehner, &

Hagerman, 2001).

Longitudinal examinations of ASD symptomatology in individuals with

FXS+ASD suggest variability in the presentation of ASD symptomatology over time. In

infancy, ASD symptoms appear to remain elevated, albeit decreasing over time (Hogan et

al., 2017), followed developmentally by reports of both increasing (Hatton et al., 2006;

Lee, Martin, Berry-Kravis, & Losh, 2016) and decreasing (Hernandes et al., 2009) ASD

symptomatology in early childhood. This variability is intriguing and may be attributable

to the wide range of measures used to assess change in ASD symptomatology across

these studies, variability present in the methods used to determine ASD diagnostic status

(i.e. parent report vs. measure cutoffs vs. clinician judgment) or developmental effects

and age related changes in symptomatology.

ASD Symptomatology in DS

Down syndrome is the most common form of ID, caused by the presence of three

copies of the 21st chromosome, in either all or a portion of the body’s cells. DS occurs at

a rate of between 1 in 400 to 1 in 1500 and affects males and females similarly (Kazemi

et al., 2016). Three types of DS are possible: trisomy 21 (95%), translocation (4%) and

mosaicism (2%). Children with the mosaicism form of DS are often less severely

impaired than the 2 other types of DS because the third copy of chromosome 21 is not

present in all cells in the body (Papavassiliou, Charalsawadi, Rafferty, & Jackson-cook,

2014).

36

Research into symptomatology and features of ASD in DS is significantly more

limited than in FXS. This may be partly attributable to the historically held belief that

individuals with DS “could not have” ASD because of the documented sociability in

many with the syndrome, with some hypothesizing that this high sociability may serve as

a protective factor against ASD for these individuals (Rasmussen, Borjesson, Elisabet, &

Gillber, 2001; Reilly, 2009). However, research has suggested that rates of ASD in DS

are elevated compared to the general population, ranging from 7-39% (Diguiseppi et al.,

2010; Hepburn, Philofsky, Fidler, & Rogers, 2008; Kent et al., 1999; Nærland, Bakke,

Storvik, Warner, & Howlin, 2017), and that the presence of comorbid DS and ASD

(DS+ASD) results in greater functional impairment (Diguiseppi et al., 2010; Dressler,

Perelli, Bozza, & Bargagna, 2011; Moore Channell et al., 2019).

Similar to young children with nsASD (Landa, 2008), research examining

individuals with DS+ASD suggests that parents report observing symptoms consistent

with ASD in infancy and toddlerhood and that their behavioral phenotype is distinct from

those with DS without ASD (DS-noASD) (Rasmussen et al., 2001). Despite symptoms

being present in infancy, most children with DS are not diagnosed with ASD until they

are between 6-14 years-of-age (Rasmussen et al., 2001) highlighting a significant missed

opportunity for targeted early intervention. Similar to findings in FXS, ASD

symptomatology may occur on a spectrum within DS, with some individuals meeting full

criteria while others show evidence of sub-threshold symptomatology. Social strengths,

including social reciprocity, using a range of facial expressions, imitation and eye-contact

have been identified in individuals with DS+ASD when contrasted to nsASD (Hepburn et

al., 2008; Starr, Berument, Tomlins, Papanikolaou, & Rutter, 2005; Warner, Howlin,

37

Salomone, Moss, & Charman, 2017). Similarly, individuals with FXS have also been

noted to display social strengths when compared to individuals with nsASD, however

their strengths are identified in the areas of social smiling, showing and directing

attention, and sharing (McDuffie et al., 2015).

Studies comparing children with DS+ASD directly to nsASD have produced

inconsistent findings. In one of the largest studies examining school age children 6-15

years old, Warner and colleagues used the clinical cutoffs on the Social Communication

Questionnaire (SCQ) to define their DS+ASD group (n=183), and identified overall

lower SCQ scores in the DS+ASD group compared the nsASD group (n=189) (Warner et

al., 2017), with the greatest strengths in the social domain including imitation skills and

gesture use. Due to methodological limitations in this study (i.e., survey data), they were

unable to control for developmental or cognitive functioning between the nsASD and

DS+ASD groups.

In contrast, Moss and colleagues who also utilized the SCQ to measure ASD

symptomatology and to create their diagnostic groups, reported no significant differences

in the symptom profiles between the DS+ASD and nsASD groups (Moss, Richards,

Nelson, & Oliver, 2013). The Moss et al., study examined a much smaller sample (i.e., 17

DS+ASD, 17 nsASD), a much broader age range (4-62 years) and matched the groups on

overall ASD symptom severity (+/- 2 points on total SCQ). Therefore, it may be the case

that when overall symptom severity between nsASD and DS+ASD are similar,

differences in social communication strengths are not measurably different.

The literature reviewed thus far relies heavily on parent report screening tools to

determine diagnostic status of ASD within their DS samples. These findings are therefore

38

limited due to possible biases in parent report of symptoms. More critically, however,

these tools are limited by their inability to take into consideration a participant’s

developmental level and how that may explain the presence or absence of certain social

communicative skills. Rather, they are blunt measures which when used with children

with ID, may not necessarily perform as designed. This is especially problematic in very

young children, or those with severe or profound ID (Starr et al., 2005).

There are only 2 published studies in DS research that utilize gold standard

diagnostic tools of ASD (e.g., Autism Diagnostic Observation Scheduled (ADOS) and

Autism Diagnostic Interview (ADI)) (Godfrey et al., 2019; Hepburn et al., 2008), with

only 1 using a clinical best estimate (CBE) procedure to examine the co-morbidity of

ASD within DS. Hepburn and colleagues (2008) assessed 20 children with DS

longitudinally, at 2 and 4 years old using a CBE procedure informed by the ADOS and

ADI, and reported a prevalence of 10%. Within the sample, 2 participants met CBE

criteria for ASD at age 2 and maintained that diagnoses at their follow up visit 2 years

later, with an observed worsening of symptoms. All participants without a co-morbid

diagnosis of ASD at their first visit, remained without ASD at their follow up

appointment. Thus, this study showed very high consistency in the diagnostic stability

across the preschool years both in those with and without diagnoses of ASD. Importantly,

the authors noted that if they had relied on ADOS scores alone, they would have

overidentified participants with DS+ASD as a number of the participants had scores on

the ADOS that fell in the ASD range but clinician view was that those features were

better accounted for by ID or other factors.

39

In a more recent study, Godfrey and colleagues (2019) matched 66 participants

with nsASD to 22 participants with DS-noASD and 11 with DS+ASD on sex, age and

verbal mental age. Participants ranged in age from 30 months to 99 months. Diagnoses of

ASD were determined by clinical cutoffs on the ADOS, not CBE, and the ADI-R was

used as the outcome variable of interest. The study produced similar findings to the

Warner et al. study (2017) in that children with ASD+DS displayed less severe

symptomatology than the nsASD group in both the social communication and restricted

and repetitive behavior domains, however more severe symptomatology than the DS-

noASD group.

Comparative Research Between FXS and DS

Currently, there are no studies comparing the development of autism

symptomatology in young children with FXS and DS. Comparative research between

FXS and DS has generally focused on language, cognitive development and adaptive

behavior suggesting unique patterns of strengths and weaknesses between the disorders

(Rice, Warren, & Betz, 2005; Valencia-naranjo & Robles-bello, 2017; Will, Caravella,

Hahn, Fidler, & Roberts, 2018). Although not directly comparing core features of ASD,

Abbeduto and colleagues compared theory of mind, a social cognitive construct often

found in deficit in individuals with ASD, in a sample of 11 to 23 year-olds with DS (n=

25) and FXS (n= 18) groupwise matched on chronological age, mental age and nonverbal

IQ (Abbeduto et al., 2001). Participants with DS were found to have greater impairment

in theory of mind functioning than individuals with FXS who performed similarly to the

typically developing control group. This finding is surprising given that rates of ASD are

40

higher in FXS than DS, and theory of mind deficits are often seen in individuals with

ASD.

Taken together, these bodies of research suggest that individuals with some forms

of syndromic ASD (i.e. FXS and DS) may evidence preservation of skills in the social

communication domain that are not typically seen in individuals with nsASD, with

possible distinct profiles between the two syndromes. It is unknown, however, when

these differences emerge or if they are consistent over time. Only one study has examined

the longitudinal development of autism symptomatology in DS, and no work has

compared the development of autism symptomatology beginning in toddlerhood, in

individuals with FXS and DS.

Comparing developmental trajectories of autism symptomatology across these

etiologically distinct, but high-risk groups, is important to determine symptom overlap

and divergence to inform the clinical phenotype of ASD in single gene disorders. Across

both bodies of work, concern exists in the potential overidentification of ASD, given

almost universal ID and possible overlap in symptoms. Indeed, concerns also exist about

the bias to underdiagnose ASD in syndromic groups, known as diagnostic

overshadowing, (Jopp & Keys, 2001), which can result in missed access to specialized

services when they are needed.

While behavioral phenotypes of DS and FXS may overlap with nsASD, evidence

suggests that identifying specific ASD phenotypes within single gene disorders is critical

to providing treatment that targets possible underlying impairments (e.g., attention,

anxiety, cognitive function, social amotivation) that give rise to the behavioral phenotype

of ASD, which may be distinct from nsASD (Glennon, Karmiloff-Smith, & Thomas,

41

2017). Understanding how these symptoms develop over time will provide important

insight into the symptom presentation of ASD in high-risk syndrome groups that can

inform screening and diagnostic practices and targets for intervention.

The Present Study

To our knowledge, there is no published literature comparing early autism

symptom trajectories in children with FXS, DS and nsASD. Given this gap in the

literature, this study aims to present preliminary findings from a pilot sample of 2-4-year-

old children with FXS, DS and nASD, measured longitudinally. While expansion of this

sample is ongoing and will be used for a complete analysis and interpretation, this initial

exploration provides insight into possible differences between groups in autism

symptomatology at 24 months. Additionally, patterns of symptom trajectories will be

examined graphically, and initial hypothesis will be drawn to inform the larger study that

will be built upon findings from this initial pilot study. Specifically, this pilot study will

report rates of clinical best estimate diagnosis of ASD in a pilot sample of males with DS

and FXS, and compare FXS, DS and nsASD groups with respect to, 1) BOSCC scores at

24 months and 2) trajectories of autism symptoms (i.e., measured by the BOSCC) from

2-4 years.

Methods

Participants

The sample included 32 male participants aged 2-4, 17 FXS, 10 DS, and 5

nsASD. Participants each had between 2 and 3 assessments, with a total of 74

assessments in the sample (38 FXS, 25. DS, 11 nsASD). Participants were drawn from a

larger longitudinal study (R01MH90194) examining the early developmental trajectories

42

of children with FXS compared to children at high and low risk of developing ASD (e.g.,

low risk controls, and infant siblings of children with ASD). A supplemental study

allowed for the addition of a pilot sample of infants with DS, to serve as a mental age

match to the FXS sample. Participants with FXS all had full mutation FXS, confirmed by

genetic report. The nsASD group includes participants from the low risk and high risk

ASIB groups diagnosed with ASD at their 24-month visit, where a genetic syndrome was

rules out. To reduce variability due to differences in cognitive functioning, all nsASD

participants were required to show developmental delays, measured by an Early Learning

Composite less than 85 on the Mullen Scales of Early Learning. The DS group was

confirmed by genetic report; 9 had Trisomy 21 and 1 had mosaicism. To be included in

the current study, participants needed to have at least 2 comprehensive diagnostic

assessments that included an ADOS, with the first occurring no later than 36 months of

age. Data was only included for assessments that occurred at ages 2, 3 and 4 years old,

given that this was the available range for the DS participants.

Exclusionary criteria for all participants included gestational age less than 37

weeks, a language other than English as the primary language in the home, and female

sex. Females were excluded because of the significant heterogeneity within the female

FXS phenotype, and the inability to account for that variability in such small pilot

samples.

Procedure

The USC Institutional Review Board approved all studies from which data will be

drawn for the present. As part of the parent study, participants were seen at 24 months for

a comprehensive diagnostic assessment for ASD. Through a renewal of the parent study,

43

participants were seen annually through the age of 5 with recurrent annual comprehensive

evaluations. Variability in the number of participants assessed at each time point occurred

based on the time between a participants’ completion of the initial study (i.e., at 24

months) and the funding of the renewal. For some, this resulted in a missed

comprehensive evaluation at 36 months (n = 6).

Visits for the parent study were primarily conducted in the participant’s homes, to

minimize travel burdens for families with children with disabilities and to reduce effects

due to novelty of the research lab. When convenient for the family, assessments were

conducted at the Neurodevelopmental Disorders Lab at USC. At each assessment a

standard battery was completed. A CBE diagnosis of ASD (Ozonoff et al., 2015) was

determined at each evaluation by a team of at least 3 members of the research staff who

were all research reliable on the ADOS, one of whom was a licensed psychologist. All

available clinical information including the Mullen Scales of Early Leaning, Vineland

Adaptive Behavior Scales and the Autism Diagnostic Observation Schedule, were used to

determine whether the child met criteria for a diagnosis of ASD.

Measures

Autism Diagnostic Observation Schedule-2. The Autism Diagnostic Observation

Schedule-2 (ADOS-2) is a standardized semi-structured play-based measure used to

observe behaviors consistent with a diagnosis of autism spectrum disorder. The ADOS-2

has 5 modules (i.e. Toddler, Modules 1-4), which are administered based on an

individual’s chronological age and expressive language level. Items are generally scored

on a scale of 0-3, with scores of 0 indicating a normative response or the absence of

atypicality, and a 3 indicating severe atypicality. Items identified as being the most

44

specific to an autism diagnosis from each module are used to create an algorithm score.

To allow for comparison across modules, algorithm scores can be converted to a

calibrated severity score, which fall on the same scale, 0-10, across all modules. The

ADOS-2 reports strong inter-rater reliability of 84% (Lord et al., 2012). Within lab inter-

rater reliability is maintained above 80%. In the present study, the ADOS was used for

two purposes. First, the ADOS was used to inform the CBE diagnosis of ASD in all

groups. Second, video clips of tasks within the ADOS were coded using the Brief

Observation of Social Communication Change (described below).

Brief Observation of Social Communication Change. The Brief Observation of

Social Communication Change (BOSCC, Grzadzinski et al., 2016) is a coding scheme

designed to measure change in autism symptomatology over time in individuals with

minimal language, when compared across 2 or more time points. The BOSCC is coded

based on 10-12 minutes of semi-structured play between an adult and the child being

assessed. The BOSCC can be applied to a range of play interactions, however they must

include developmentally appropriate play and toy selection. The child must also be

allowed to freely move around the room and explore toys. Additionally, whatever setting

and structure is selected for the initial observation should be kept consistent for all future

play sessions to which the BOSCC will be coded and compared. Fifteen decision trees

are used to derive codes based on behaviors observed during the session. Observations

are split into two 6-minute segments, which are coded independently then averaged to

determine a child’s total score on the BOSCC. For the present study, the BOSCC coding

scheme will be applied to video clips selected from the participant’s ADOS-2. Video

clips will include the following tasks from the ADOS-2 administrations; 3 minutes of

45

Free Play, 3 minutes of Bubble Play, 3 Minutes of Birthday Party or Bath Time and 3

Minutes of Balloon Play. This collection of tasks and time lengths is recommended by

the authors and has been previously published on by Kitzerow and colleagues (Kitzerow,

Teufel, Wilker, & Freitag, 2016). The first author of the current study has attended

research training on the coding of the BOSCC and has been given authorization to use it

for research purposes by the authors. Consistent with measure use guidelines, a team of

undergraduate research assistants were trained to reliability standards with the first

author. To achieve reliability with the first author, coders must score within 1 point for

more than 80% of items on each coding segment and within 4 points on segment total

scores. This requirement was met for 3 consecutive videos before coders were able to

code independently. Additionally, approximately 20% of videos were coded for

reliability to minimize coding drift. Intraclass correlations were used to examine

reliability of absolute agreement, which was high (ICC = .911).

Mullen Scales of Early Learning. The Mullen Scales of Early Learning (Mullen,

(Mullen, 1995) is a standardized measure of development normed for young children

ages 0-60 months. The Mullen measures development across five domains; Gross Motor,

Fine Motor, Receptive Language, Expressive Language, and Visual Reception. A

summary standard score, the early learning composite (ELC) represents a composite

score across all domains, with the exclusion of gross motor. The ELC has a mean of 100,

and standard deviation of 15. The Mullen was administered by trained research staff.

Internal consistency for each of the skills ranges from 0.75 to .08, and coefficients of test-

retest reliability range from .70- .80 (Mullen, 1995). The Mullen ELC was used as

inclusionary for the nsASD group (i.e., nsASD participants needed to have an ELC < 85.)

46

Data Analytic Plan

Given small samples, descriptive analysis and simple comparisons were utilized

in lieu of complex inferential models. BOSCC scores from the first time point were

compared across groups using an ANOVA and Tukey HSD post hoc tests. Tukey HSD

tests analyze all possible pairwise comparisons and provides p-values that are corrected

for multiple comparisons. Comparisons at 36 and 48 months were not conducted; due to

the smaller samples at 36 and 48 months, a repeated measures ANOVA was not

appropriate. Third, BOSCC change scores were calculated as: Last BOSCC Total – First

BOSCC Total / Time in Months Between Assessments. One sample T-tests were

conducted to determine if change scores differed from 0.

Results

Clinical Diagnoses of Autism within FXS and DS participants

Within the FXS group, 65% of participants (n=11) met CBE criteria for ASD.

One participant was a false negative for ASD at his 24-month visit and was diagnosed

with ASD at both subsequent visits (36 and 48 months). Therefore, he is included in the

FXS+ASD group for all comparisons. All other participants displayed stable diagnoses of

either FXS+ASD or FXS-noASD throughout the duration of their participation in the

study.

Within the DS group, 20% of participants (n = 2) met CBE criteria for ASD.

These two participants were false negatives at their 24-month visit and diagnosed with

ASD at subsequent visits. The remaining participants did not meet criteria for ASD at any

time during their participation in the study.

47

Examination of BOSCC Means and Trajectories

All but 1 participant had their first available BOSCC at 24 months. Means,

standard deviations and sample sizes at each age point are presented in Table 2.1. A

Shapiro-Wilk’s tests of normality was conducted on BOSCC scores at each time point

and determined that BOSCC score distributions did not differ from normality at any time

point. An ANOVA comparing BOSCC means at 24 months identified differences across

groups, F(4,26) = 4.94, p = .004. Post hoc comparisons using a Tukey HSD test, which

produce p-values corrected for multiple comparisons, indicated that the FXS+ASD (M =

29.68) group scored higher than the DS-noASD group (M = 17.86, p = .021) and higher

than the FXS-noASD group (M = 15.08, p = .005) on BOSCC scores at 24 months. The

average BOSCC score in the nsASD group at 24 months was 25.6, which is not

statistically different than any of the comparison groups. The DS+ASD group had a

mean BOSCC score of 22, which was the lowest of all groups with a co-morbid diagnosis

of ASD, although this was not determined to be statistically different from any

comparison group.

Longitudinal trajectories of BOSCC data by group are presented in Figure 2.1.

Change scores for each group, which represented the average change in BOSCC score

per month, are presented in Table 2.1. One sample T-tests were used to compare change

scores to 0. Results of the T-tests suggest that change trajectories for each group are not

different than 0 (i.e., all groups evidence flat BOSCC trajectories). However, these results

are likely affected by the small sample sizes. Visual examination of individual growth

trajectories highlights the heterogeneity within each group, which may be not best

captured by group level means.

48

Discussion

This study reports findings from a pilot cross-syndrome longitudinal study of

ASD symptom trajectories in toddlers and preschoolers with DS, FXS or nsASD. This

study has two primary contributions. First, these pilot data provide important information

to guide the design of a more systematic and large-scale study with additional

participants. Second, given that no study has been published to contrast ASD trajectories

across these syndromes, the findings are important to the field. Given the pilot nature of

this study, caution is taken with regard to conclusions and inferences with limitations on

generalizability recognized. As such, the patterns observed from the DS+ASD group will

be described, however conclusive inferences will not be made given the sample size.

At the diagnostic level, rates of syndromic ASD in this pilot sample were

consistent with rates reported in the literature across both the FXS and DS groups, at 65%

and 20% respectively. Interestingly, both participants in the DS sample were false

negatives for ASD at their 24 month visit and were not diagnosed with ASD until their 36

month assessment. In the only other study to conduct comprehensive assessments for

ASD longitudinally in the toddler years in a DS sample, both participants with DS+ASD

(n = 2) met criteria for ASD at 24 months and maintained this diagnosis over time

(Hepburn et al., 2008). Future research should aim to identify what factors (e.g.,

individual level vs. clinician factors) may lead to these patterns of ASD diagnostic

instability in DS, as they have significant implications for clinical practice in screening

and surveillance.

This study suggests that by 24 months, participants in the FXS+ASD group are

scoring highest on the BOSCC, with scores that are significantly higher than individuals

49

in both syndrome groups without ASD (i.e., FXS-noASD and DS-noASD). This finding

aligns with the theoretical approach of a specific ASD phenotype within FXS, that is

distinct from FXS without co-morbid ASD. In contrast, the FXS+ASD group could not

be differentiated at 24 months from the DS+ASD and nsASD groups, suggesting that on

average, these groups show similar levels of autism symptomatology at 24 months. This

finding, if replicated in a larger sample, would suggest that toddlers with syndromic ASD

and nsASD show similar composite levels of autism symptomatology during the toddler

years. This lack of differentiation in autism symptomatology between FXS+ASD and

nsASD groups in the toddler years is consistent with previous research (Hazlett, Poe,

Lightbody, Gerig, MacFall, et al., 2009; Rogers et al., 2001).

Although participants with DS+ASD did not significantly differ from the nsASD

group, they also did not differ significantly from the DS-noASD group. The DS+ASD

group in this study display an intermediate autism symptom phenotype that falls between

the nsASD and DS-noASD group (i.e., nsASD > DS+ASD > DS-noASD), although these

differences did not rise to levels of statistical significance. However, this ASD symptom

severity gradient within DS is consistent with what has been observed in larger studies

(Godfrey et al., 2019; Warner et al., 2017). In contrast to the DS sample, participants in

the FXS sample were distinguishable by almost 16 points on the BOSCC at their 24-

month time point, based on their clinical diagnoses (i.e., FXS+ASD vs. FXS-noASD).

This difference suggests there may be a much stronger distinction between ASD

diagnostic groups within FXS, with more subtle differentiations in individuals with DS.

This possible contrasting finding between the two syndrome groups will need to be

replicated in a much larger sample.

50

At conception, this study was focused on comparing ASD symptom change

trajectories across time between non-syndromic and syndromic forms of ASD. The

BOSCC was utilized due to its ability to capture small and clinically relevant changes

over a short period of time. In the present study, slopes across all groups could not be

differentiated from 0 at the mean level. However, visual examination of BOSCC

trajectories (Figure 2.1) highlights the variability in slopes that are most often non-zero.

At a sample level, group does appear to explain variance in total BOSCC scores,

evidenced by the descending pattern in mean trend lines plotted by group in Figure 2.1.

However, the within group variability in slopes is masked at the mean level. Identifying

predictors of within group variability will be essential to explaining and interpreting this

heterogeneity.

Limitations and Future Directions

While this pilot study presents a novel comparison of ASD symptom trajectories

in 2-4-year olds with nsASD, FXS and DS, it is limited by the small samples, and

therefore generalizability is not yet possible. Since the inception of this pilot study, data

collection across all three groups has continued through the parent study. Therefore, a

follow up analysis is planned on a larger sample of participants, that is projected to

include: 25 participants with nsASD with developmental delay, 16 participants with DS,

and 21 FXS. The expanded sample will take advantage of multilevel models of change,

which are robust to uneven time points, which are present in this dataset.

Prospective research interested in examining the development of autism

symptomatology in young children with DS is inherently challenged by the relatively low

prevalence rates of ASD within DS. With reported prevalence estimates ranging from 7-

51

39%, studies aiming for an outcome group of participants with DS+ASD need to enroll

somewhere between 65 – 350 participants, to end up with a modest sample size of 25

participants. This sample size challenge is evident in the two studies that used gold

standard in person diagnostic assessment of ASD, which included 2 (Hepburn et al.,

2008) and 11 participants (Godfrey et al., 2019) with DS+ASD respectively.

This recruitment challenge is shared by researchers who prospectively study

autism infant siblings at risk for developing ASD (ASIBs), where prevalence rates of

ASD are similar, at approximately 20% (Ozonoff et al., 2011). To address this challenge

in ASIB research, collaboration across sites and consortia (e.g., Baby Siblings Research

Consortium, Infant Brain Imaging Study) are capitalized upon to produce large samples

of children with outcome diagnoses of ASD. A recently launched NIH-Wide initiative

“INCLUDE (Investigation of Co-occurring conditions across the Lifespan to Understand

Down syndromE)”, lists ASD as a critical co-occurring condition that will be a target of

this initiative. Hopefully, through targeted funding and collaborative research efforts,

larger samples can be attained so that these important questions can be investigated to

guide screening, diagnostic and treatment efforts to improve outcomes for individuals

with DS and comorbid ASD.

52

Table 2.1 Mean BOSCC scores by group across time

FXS

DS

Non-Syndromic

FXS+ASD n = 11

FXS-noASD

n = 6

DS+ASD

n = 2

DS-noASD

n =8

nsASD n = 5

24 months 29.68 (9.42) *+ 15.08 (6.76)* 22.00 (4.24) 17.86 (5.77)+ (n = 7)

25.6 (4.97)

36 months 32.71 (7.99) (n = 7)

12.6 (4.85) (n = 5)

21.25 (4.60) (n = 2)

15.25 (5.94) (n = 8)

26.75 (7.44) (n = 4)

48 months 33.5 (7.26) (n = 7)

13.75 (5.30) (n = 2)

23.5 (--) (n = 1)

15.1 (8.93) (n = 5)

20.5 (9.90) (n = 2)

Average change score per month

0.23 (0.50) -0.05 (0.35) -0.24 (0.62) -0.07 (0.38) -0.08 (0.35)

*FXS+ASD > FXS-noASD + FXS+ASD > DS-noASD FXS, Fragile X syndrome; DS, Down syndrome; ASD, autism spectrum disorder; nsASD, non-syndromic autism spectrum disorder; BOSCC, Brief Observation of Social Communication Change

53

REFERENCESFigure 2.1 Individual spaghetti plots of BOSCC total score trajectories between groups

54

REFERENCES

Abbeduto, L., Brady, N. C., & Kover, S. T. (2007). Language Development and Fragile X Syndrome: Progiles, Syndrome-specificity and within-syndrome differences. Mental Retardation and Developmental Disabilities Research Reviews, 13, 36–46. https://doi.org/10.1002/mrdd

Abbeduto, L., McDuffie, A., & Thurman, A. J. (2014). The fragile X syndrome-autism comorbidity: what do we really know? Frontiers in Genetics, 5(October), 355. https://doi.org/10.3389/fgene.2014.00355

Abbeduto, L., Pavetto, M., Kesin, E., Weissman, M. D., Karadottir, S., Brien, A. O., & Cawthon, S. (2001). The linguistic and cognitive profile of Down syndrome : Evidence from a comparison with fragile X syndrome. Downs Syndrome Research

and Practice, 7(1), 9–15.

Abbeduto, L., Thurman, A. J., McDuffie, A., Klusek, J., Feigles, R. T., Ted Brown, W., … Roberts, J. E. (2019). ASD Comorbidity in Fragile X Syndrome: Symptom Profile and Predictors of Symptom Severity in Adolescent and Young Adult Males. Journal of Autism and Developmental Disorders, 49(3), 960–977. https://doi.org/10.1007/s10803-018-3796-2

Abrahams, B. S., & Geschwind, D. H. (2008). Advances in autism genetics: on the threshold of a new neurobiology. Nature Reviews. Genetics, 9(5), 341–355. https://doi.org/10.1038/nrg2346

American Psychiatric Association. (2013). The Diagnostic and Statistical Manual of

Mental Disorders: DSM 5. bookpointUS.

Anagnostou, E., Jones, N., Huerta, M., Halladay, A. K., Wang, P., Scahill, L., … Dawson, G. (2015). Measuring social communication behaviors as a treatment endpoint in individuals with autism spectrum disorder. Autism, 19(5), 622–636. https://doi.org/10.1177/1362361314542955

Association, A. P. (2013). DSM 5. American Psychiatric Association.

55

Bailey, D. B., Hatton, D. D., Mesibov, G., Ament, N., & Skinner, M. (2000). Early development, temperament, and functional impairment in autism and fragile X syndrome. Journal of Autism and Developmental Disorders, 30(1), 49–59. https://doi.org/10.1023/A:1005412111706

Bailey, D. B., Hatton, D. D., Tassone, F., Skinner, M., & Taylor, A. K. (2001). Variability in FMRP and early development in males with fragile X syndrome. American Journal of Mental Retardation, 106(1), 16–27. Bailey, Donald B, Hatton, D. D., Skinner, M., & Mesibov, G. (2001). Autistic Behavior , FMR1 Protein , and Developmental Trajectories in Young Males with Fragile X Syndrome. Journal of

Autism and Developmental Disorders, 31(2), 165–174.

Bryson, S. E., Zwaigenbaum, L., McDermott, C., Rombough, V., & Brian, J. (2008). The Autism Observation Scale for Infants: scale development and reliability data. Journal of Autism and Developmental Disorders, 38(4), 731–738. https://doi.org/10.1007/s10803-007-0440-y

Caravella, K. E., & Roberts, J. E. (2017). Adaptive skill trajectories in infants with fragile X syndrome contrasted to typical controls and infants at high risk for autism. Research in Autism Spectrum Disorders, 40. https://doi.org/10.1016/j.rasd.2017.05.002

Charman, T., Pickles, A., Simonoff, E., Chandler, S., Loucas, T., & Baird, G. (2011). IQ in children with autism spectrum disorders: data from the Special Needs and Autism Project (SNAP). Psychological Medicine, 41(3), 619–627. https://doi.org/10.1017/S0033291710000991

Chawarska, K., Klin, A., Paul, R., & Volkmar, F. (2007). Autism spectrum disorder in the second year: stability and change in syndrome expression. Journal of Child

Psychology and Psychiatry, and Allied Disciplines, 48(2), 128–138. https://doi.org/10.1111/j.1469-7610.2006.01685.x

Coffee, B., Keith, K., Albizua, I., Malone, T., Mowrey, J., Sherman, S. L., & Warren, S. T. (2009). Incidence of Fragile X Syndrome by Newborn Screening for Methylated FMR1 DNA. American Journal of Human Genetics, 85(4), 503–514. https://doi.org/10.1016/j.ajhg.2009.09.007

Cohen, I. L., Sudhalter, V., Pfadt, a, Jenkins, E. C., Brown, W. T., & Vietze, P. M. (1991). Why are autism and the fragile-X syndrome associated? Conceptual and methodological issues. American Journal of Human Genetics, 48(2), 195–202.

56

Control, C. for D. (2014). Prevalence of autism spectrum disorder among children aged 8 years - autism and developmental disabilities monitoring network, 11 sites, United States, 2010. MMWR. Surveillance Summaries, 63(2), 1–21. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/24670961

DiGuiseppi, C., Hepburn, S., Davis, J. M., Fidler, D. J., Hartway, S., Lee, N. R., … Robinson, C. (2010). Screening for Autism Spectrum Disorders in Children With Down Syndrome. Journal of Developmental & Behavioral Pediatrics, 31(3), 181–191. https://doi.org/10.1097/DBP.0b013e3181d5aa6d

Dressler, A., Perelli, V., Bozza, M., & Bargagna, S. (2011). The autistic phenotype in Down syndrome: Differences in adaptive behaviour versus Down syndrome alone and autistic disorder alone. Functional Neurology, 26(3), 151–158.

Garber, K. B., Visootsak, J., & Warren, S. T. (2008). Fragile X syndrome. European

Journal of Human Genetics, 16(6), 666–672. https://doi.org/10.1038/ejhg.2008.61

Geschwind, D. H. (2011). Genetics of autism spectrum disorders. Trends in Cognitive

Sciences, 15(9), 409–416. https://doi.org/10.1016/j.tics.2011.07.003

Geschwind, D. H., & Levitt, P. (2007). Autism spectrum disorders : developmental disconnection syndromes. Neurobiology, (17), 103–111. https://doi.org/10.1016/j.conb.2007.01.009

Glennon, J. M., Karmiloff-Smith, A., & Thomas, M. S. C. (2017). Syndromic Autism: Progressing Beyond Current Levels of Description. Review Journal of Autism and

Developmental Disorders, 4(4), 321–327. https://doi.org/10.1007/s40489-017-0116-2

Godfrey, M., Hepburn, S., Fidler, D. J., Tapera, T., Zhang, F., Rosenberg, C. R., & Raitano Lee, N. (2019). Autism spectrum disorder (ASD) symptom profiles of children with comorbid Down syndrome (DS) and ASD: A comparison with children with DS-only and ASD-only. Research in Developmental Disabilities, 89(October 2018), 83–93. https://doi.org/10.1016/j.ridd.2019.03.003

Grzadzinski, R., Carr, T., Colombi, C., McGuire, K., Dufek, S., Pickles, A., & Lord, C. (2016). Measuring Changes in Social Communication Behaviors: Preliminary Development of the Brief Observation of Social Communication Change (BOSCC). Journal of Autism and Developmental Disorders, 46(7), 2464–2479. https://doi.org/10.1007/s10803-016-2782-9

57

Guthrie, W., Swineford, L. B., Nottke, C., & Wetherby, A. M. (2013). Early diagnosis of autism spectrum disorder: stability and change in clinical diagnosis and symptom presentation. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 54(5), 582–590. https://doi.org/10.1111/jcpp.12008

Hagerman, R. J., & Hagerman, P. J. (2002). Fragile X syndrome: Diagnosis, treatment,

and research. JHU Press.

Hahn, L. J., Brady, N. C., Warren, S. F., & Fleming, K. K. (2015). Do Children With Fragile X Syndrome Show Declines or Plateaus in Adaptive Behavior? American

Journal on Intellectual and Developmental Disabilities, 120(5), 412–432. https://doi.org/10.1352/1944-7558-120.5.412

Hall, S. S., Lightbody, A. A., Hirt, M., Rezvani, A., & Reiss, A. L. (2010). Autism in Fragile X Syndrome : A Category Mistake ? Journal of the American Academy of

Child & Adolescent Psychiatry, 49(9), 921–933.

Harris, S. W., Hessl, D., Goodlin-Jones, B., Ferranti, J., Bacalman, S., Barbato, I., … Hagerman, R. J. (2008). Autism profiles of males with fragile X syndrome. Am J

Ment Retard, 113(6), 427–438.

Hatton, D. D., Sideris, J., Skinner, M., Mankowski, J., Bailey, D. B., Roberts, J., & Mirrett, P. (2006). Autistic behavior in children with fragile X syndrome: Prevalence, stability, and the impact of FMRP. American Journal of Medical

Genetics Part A, 140A(17), 1804–1813. https://doi.org/10.1002/ajmg.a.31286

Hazlett, H. C., Poe, M. D., Lightbody, A. A., Gerig, G., MacFall, J. R., Ross, A. K., … Piven, J. (2009). Teasing apart the heterogeneity of autism: Same behavior, different brains in toddlers with fragile X syndrome and autism. Journal of

Neurodevelopmental Disorders, 1(1), 81–90. https://doi.org/10.1007/s11689-009-9009-8

Hepburn, S., Philofsky, A., Fidler, D. J., & Rogers, S. (2008). Autism symptoms in toddlers with Down syndrome: A descriptive study. Journal of Applied Research in

Intellectual Disabilities, 21(1), 48–57. https://doi.org/10.1111/j.1468-3148.2007.00368.x

Hernandez, R. N., Feinberg, R. L., Vaurio, R., Passanante, N. M., Thompson, R. E., & Kaufmann, W. E. (2009). Autism spectrum disorder in fragile X syndrome: A longitudinal evaluation. American Journal of Medical Genetics Part A, 149A(6), 1125–1137. https://doi.org/10.1002/ajmg.a.32848

58

Hogan, A. L., Caravella, K. E., Ezell, J., Rague, L., Hills, K., & Roberts, J. E. (2017). Autism Spectrum Disorder Symptoms in Infants with Fragile X Syndrome: A Prospective Case Series. Journal of Autism and Developmental Disorders, 47(6), 1628–1644. https://doi.org/10.1007/s10803-017-3081-9

Jopp, D. A., & Keys, C. B. (2001). Diagnostic Overshadowin Reviewed and Reconsidered. American Journal on Mental Retardation, 106(5), 416–433.

Kaufmann, W. E., Cortell, R., Kau, A. S. M., Bukelis, I., Tierney, E., Gray, R. M., … Stanard, P. (2004). Autism spectrum disorder in fragile X syndrome : Communication , social interaction , and specific behaviors. American Journal of

Medical Genetics, 129A(July), 225–234. https://doi.org/10.1002/ajmg.a.30229

Kazemi, M., Salehi, M., & Kheirollahi, M. (2016). Down Syndrome: Current Status, Challenges and Future Perspectives. International Journal of Molecular and

Cellular Medicine, 5(3), 125–133.

Kent, L., Clinic, P., Evans, J., Registrar, S., Centre, O., & Oak, S. (1999). Comorbidity of autistic spectrum disorders in children with Down syndrome. Developmental

Medicine and Child Neurology, 41, 153–158.

Kitzerow, J., Teufel, K., Wilker, C., & Freitag, C. M. (2016). Using the brief observation of social communication change (BOSCC) to measure autism-specific development. Autism Research, 9(9), 940–950. https://doi.org/10.1002/aur.1588

Klusek, J., Martin, G. E., & Losh, M. (2014). Consistency between research and clinical diagnoses of autism among boys and girls with fragile X syndrome. Journal of

Intellectual Disability Research, 58(10), 940–952. https://doi.org/10.1111/jir.12121

Landa, R. J. (2008). Diagnosis of autism spectrum disorders in the first 3 years of life. Nature Clinical Practice. Neurology, 4(3), 138–147. https://doi.org/10.1038/ncpneuro0731

Lee, M., Martin, G. E., Berry-Kravis, E., & Losh, M. (2016). A developmental, longitudinal investigation of autism phenotypic profiles in fragile X syndrome. Journal of Neurodevelopmental Disorders, 8(1), 47. https://doi.org/10.1186/s11689-016-9179-0

Lord, C., Rutter, M., DiLavore, P. C., Risi, S., Gotham, K., & Bishop, S. L. (2012). Autism Diagnostic Observation Schedule, Second Edition (ADOS-2). Los Angeles, CA: Western Psychological Services.

59

Martin, G. E., Ausderau, K. K., Raspa, M., Bishop, E., Mallya, U., & Jr, D. B. B. (2013). Therapy service use among individuals with fragile X syndrome : findings from a US parent survey. Journal of Intellectual Disability Research, 57(9), 837–849. https://doi.org/10.1111/j.1365-2788.2012.01608.x

Matson, J. L., & Shoemaker, M. (2009). Intellectual disability and its relationship to autism spectrum disorders. Research in Developmental Disabilities, 30(6), 1107–1114. https://doi.org/10.1016/j.ridd.2009.06.003

McDuffie, A., Thurman, A. J., Hagerman, R. J., & Abbeduto, L. (2014). Symptoms of Autism in Males with Fragile X Syndrome: A Comparison to Nonsyndromic ASD Using Current ADI-R Scores. Journal of Autism and Developmental Disorders, 1–13. https://doi.org/10.1007/s10803-013-2013-6

McDuffie, A., Thurman, A. J., Hagerman, R. J., & Abbeduto, L. (2015). Symptoms of Autism in Males with Fragile X Syndrome: A Comparison to Nonsyndromic ASD Using Current ADI-R Scores. Journal of Autism and Developmental Disorders, 45(7), 1925–1937. https://doi.org/10.1007/s10803-013-2013-6

Moore Channell, M., Hahn, L. J., Rosser, T. C., Hamilton, D., Frank-Crawford, M. A., Capone, G. T., & Sherman, S. L. (2019). Characteristics Associated with Autism Spectrum Disorder Risk in Individuals with Down Syndrome. Journal of Autism and

Developmental Disorders, (0123456789). https://doi.org/10.1007/s10803-019-04074-1

Moss, J., Richards, C., Nelson, L., & Oliver, C. (2013). Prevalence of autism spectrum disorder symptomatology and related behavioural characteristics in individuals with Down syndrome. Autism, 17(4), 390–404. https://doi.org/10.1177/1362361312442790

Mullen, E. (1995). Mullen scales of early learning. Circle Pines, MN: American Guidance Service. Inc.(21).

Nærland, T., Bakke, K. A., Storvik, S., Warner, G., & Howlin, P. (2017). Age and gender-related differences in emotional and behavioural problems and autistic features in children and adolescents with Down syndrome: a survey-based study of 674 individuals. Journal of Intellectual Disability Research, 61(6), 594–603. https://doi.org/10.1111/jir.12342

Ozonoff, S., Young, G. S., Carter, A., Messinger, D., Yirmiya, N., Zwaigenbaum, L., … Stone, W. L. (2011). Recurrence risk for autism spectrum disorders: a Baby Siblings Research Consortium study. Pediatrics, 128(3), e488-95. https://doi.org/10.1542/peds.2010-2825

60

Ozonoff, S., Young, G. S., Landa, R. J., Brian, J., Bryson, S., Charman, T., … Iosif, A. (2015). Diagnostic stability in young children at risk for autism spectrum disorder: a baby siblings research consortium study. Journal of Child Psychology and

Psychiatry, 56(9), 988–998. https://doi.org/10.1111/jcpp.12421.Diagnostic

Papavassiliou, P., Charalsawadi, C., Rafferty, K., & Jackson-Cook, C. (2014). Mosaicism for Trisomy 21 : A Review. American Journal of Medical Genetics, (November), 26–39. https://doi.org/10.1002/ajmg.a.36861

Rasmussen, P., Borjesson, O., Elisabet, W., & Gillberg, C. (2001). Autistic disorders in Down syndrome : background factors and clinical correlates. Developmental

Medicine & Child Neurology, 43, 750–754. https://doi.org/10.1017/S0012162201001372

Reilly, C. (2009). Autism spectrum disorders in Down syndrome: A review. Research in

Autism Spectrum Disorders, 3(4), 829–839. https://doi.org/10.1016/j.rasd.2009.01.012

Rice, M. L., Warren, S. F., & Betz, S. K. (2005). Language symptoms of developmental language disorders: An overview of autism, Down syndrome, fragile X, specific language impairment, and Williams syndrome. Applied Psycholinguistics, 26(1), 7–27. https://doi.org/10.1017.S0142716405050034

Rivard, M. ´lina, Terroux, A. ´lie, Mercier, C. ’line, & Parent-Boursier, C. (2015). Indicators of Intellectual Disabilities in Young Children with Autism Spectrum Disorders. Journal of Autism and Developmental Disorders, 45, 127–137. https://doi.org/10.1007/s10803-014-2198-3

Roberts, J., Crawford, H., Hogan, A. L., Fairchild, A., Tonnsen, B., Brewe, A., … Abbeduto, L. (2019). Social Avoidance Emerges in Infancy and Persists into Adulthood in Fragile X Syndrome. Journal of Autism and Developmental Disorders, (0123456789). https://doi.org/10.1007/s10803-019-04051-8

Roberts, J. E., Tonnsen, B. L., McCary, L. M., Caravella, K. E., & Shinkareva, S. V. (2016). Brief Report: Autism Symptoms in Infants with Fragile X Syndrome. Journal of Autism and Developmental Disorders, 1–8. https://doi.org/10.1007/s10803-016-2903-5

Roberts, J. E., Weisenfeld, L. A. H., Hatton, D. D., Heath, M., & Kaufmann, W. E. (2007). Social approach and autistic behavior in children with fragile X syndrome. Journal of Autism and Developmental Disorders, 37(9), 1748–1760.

61

Rogers, S. J., Wehner, E. A., & Hagerman, R. (2001). The Behavioral Phenotype in Fragile X: Symptoms of Autism in Very Young Chilren with Fragile X Syndrome, Idiopathic Autism and Other Developmental Disorders. Developmetnal and

Behavioral Pediatrics, 22(6), 409–417.

Schopler, E., Reichler, R. J., & Renner, B. R. (2002). The childhood autism rating scale

(CARS). Western Psychological Services Los Angeles, CA.

Singer, J. D., Willett, J. B., & Willett, J. B. (2003). Applied longitudinal data analysis:

Modeling change and event occurrence. Oxford university press.

Starr, E. M., Berument, S. K., Tomlins, M., Papanikolaou, K., & Rutter, M. (2005). Brief Report : Autism in Individuals with Down Syndrome Brief Report : Autism in Individuals with Down Syndrome. Journal of Autism and Developmental Disorders, 35(5), 665–673. https://doi.org/10.1007/s10803-005-0010-0

Valencia-naranjo, N., & Robles-bello, M. A. (2017). Research in Developmental Disabilities Learning potential and cognitive abilities in preschool boys with fragile X and Down syndrome. Research in Developmental Disabilities, 60, 153–161. https://doi.org/10.1016/j.ridd.2016.12.001

Vorstman, J. A. S., Spooren, W., Persico, A. M., Collier, D. A., Aigner, S., Jagasia, R., … Buitelaar, J. K. (2013). Using genetic findings in autism for the development of new pharmaceutical compounds. Psychopharmacology. https://doi.org/10.1007/s00213-013-3334-z

Wallander, J. L., Dekker, M. C., & Koot, H. M. (2003). Psychopathology in Children and Adolescents with Intellectual Disability: Measurement, Prevalence, Course, and Risk. International Review of Research in Mental Retardation, 26, 93–134. https://doi.org/10.1016/S0074-7750(03)01003-6

Warner, G., Howlin, P., Salomone, E., Moss, J., & Charman, T. (2017). Profiles of children with Down syndrome who meet screening criteria for autism spectrum disorder (ASD): a comparison with children diagnosed with ASD attending specialist schools. Journal of Intellectual Disability Research, 61(1), 75–82. https://doi.org/10.1111/jir.12344

Will, E. A., Caravella, K. E., Hahn, L. J., Fidler, D. J., & Roberts, J. E. (2018). Adaptive behavior in infants and toddlers with Down syndrome and fragile X syndrome. American Journal of Medical Genetics, Part B: Neuropsychiatric Genetics. https://doi.org/10.1002/ajmg.b.32619

62

Wolff, J. J., Bodfish, J. W., Hazlett, H. C., Lightbody, A. A., Reiss, A. L., Piven, J., & Ph, D. (2012). Evidence of a distinct behavioral phenotype in young boys with fragile X syndrome and autism. Journal of the American Academy of Child and

Adolescent Psychiatry, 51(12), 1324–1332. https://doi.org/10.1016/j.jaac.2012.09.001

Woolfenden, S., Sarkozy, V., Ridley, G., & Williams, K. (2012). A systematic review of the diagnostic stability of Autism Spectrum Disorder. Research in Autism Spectrum

Disorders, 6(1), 345–354. https://doi.org/10.1016/j.rasd.2011.06.008